The Two Positively Acting Regulatory Proteins PH02 and PH04 Physically

MOLECULAR AND CELLULAR BIOLOGY, May 1989, p. 2050-2057 Vol. 9, No. 5 0270-7306/89/052050-08$02.00/0 Copyright © 1989, American Society for Microbiology The Two Positively Acting Regulatory Proteins PH02 and PH04 Physically Interact with PHOS Upstream Activation Regions KURT VOGEL,1 WOLFRAM HORZ,2 AND ALBERT HINNENl* Biotechnology Department, CIBA-GEIGY AG, CH4002 Basel, Switzerland,' and Institut fur Physiologische Chemie, Physikalische Biochemie, und Zellbiologie der Universitdt Munchen, 8000 Munich 2, Federal Republic ofGermany Received 12 September 1988/Accepted 7 February 1989

The repressible acid phosphatase gene PHOS of Saccharomyces cerevisiae requires the two positively acting regulatory proteins PHO2 and PHO4 for expression. pho2 or pho4 mutants are not able to derepress the PHOS gene under low-Pi conditions. Here we show that both PHO2 and PHO4 bind specifically to the PHOS promoter in vitro. Gel retardation assays using promoter deletions revealed two regions involved in PH04 binding. Further characterization by DNase I footprinting showed two protected areas, one located at -347 to -373 (relative to the ATG initiator codon) (UASpl) and the other located at -239 to -262 (UASp2). Exonuclease m footprint experiments revealed stops at -349 and -368 (UASpl) as well as at -245 and -260 (UASp2). Gel retardation assays with the PH02 protein revealed a binding region that lay between the two PH04-binding sites. DNase I footprint analysis suggested a PH02-binding site covering the region between -277 and -296.

The promoter region of the strongly regulated acid phos- press). Interestingly, the pho2 mutation showed no influence phatase gene PH05 (4) has been recently analyzed by on the basal level of TRP4 expression but rather modulated deletion mapping (6, 19, 24). In addition to the TATA box, the general control response of TRP4. two regions were found (24) to be essential for activation of the PHOS promoter. Sequence comparisons defined four MATERIALS AND METHODS elements from which a 19-base-pair dyad consensus sequence was deduced. Three of the elements, called UASp, Escherichia coli vector for PH04 and PH02 expression. The were located within deletions showing reduced transcription PH04 gene (13, 15) was reisolated from a S288C background of PHOS (24), at -367, -245, and -185 relative to the by screening an ordered centromere library (25). An NcoI translational start. site at the ATG of PH04 was introduced by site-specific Not much is known about the physical interaction of mutagenesis (14). The PH04 gene (NcoI-BamHI fragment) regulatory proteins with the PHOS promoter. Since muta- was placed under control of the inducible PL promoter by tions defective in PH02 or PH04 are epistatic to all other using the vector pPLmu, a derivative of PLc24 (23) contain- regulatory mutations and show a negative phenotype, it is ing the phage Mu Ner gene ribosome-binding site with an likely that the PHO2 and PHO4 proteins are positive tran- NcoI site at its ATG start codon (obtained from Biogen). scription factors that interact with the PHOS promoter The PH02 gene (25) was fused to the same vector but DNA. The finding that a pho2 mutation can be comple- carrying a different ribosome-binding site (no. 27-4878-01 mented by overexpression of the PHO4 protein (20) is a and 27-4898-01; Pharmacia, Uppsala, Sweden; vector kindly further indication that the PHO2 and PHO4 proteins exert provided by A. Schmitz). The correct fusion of this ribo- their functions at the same late hierarchical level of PH05 some-binding site to the PH02 gene was achieved with an activation. adapter oligonucleotide that fits to the BamHI site of the Extensive studies of the chromatin fine structure of the ribosome-binding site and the EcoRI site of the PH02 gene. PH05 gene (1, 2, 5, 7) have been done. Interestingly, the The correct in-frame fusions of both constructions were element at -367, which is essential for gene induction (24), verified by sequencing the junctions. resides in the center of a hypersensitive region under condi- Partial purification of E. coli-expressed PH04 and PH02 tions of PH05 repression (1, 2). proteins. The E. coli expression vector with the PH04 or In this paper, we show that the two positive regulatory PH02 genes was transformed into JM109 cells carrying the proteins PHO4 and PHO2 bind specifically to PH05 pro- temperature-sensitive X repressor pcI857 on a plasmid. moter DNA. Footprint analysis revealed two regions Upon heat induction, heterologous protein accumulated to a (UASp1 and UASp2) protected by the PHO4 protein. The few percent of the level of total cellular protein. For binding PH05 promoter was found to have a PHO2-binding site studies, a soluble cell lysate (1.2 g of total protein for PHO4 located between the two PHO4-binding sites. These results and 0.5 g for PHO2) was partially purified by applying it to a are in agreement with data obtained by our fine-structure DE-52 column (30-ml bed volume) in 100 mM Tris (pH deletion mapping (24) but differ to some extent from data 7.5)-0.1 mM EDTA-1 mM phenylmethylsulfonyl fluoride published by Nakao et al. (19) and Bergman et al. (6). and rinsing the preparation with the same buffer. The Recently, it was shown that BAS2 (allelic to PHO2) also flowthrough fractions were passed through an Affi-Gel- binds to the HIS4 promoter region (3) and regulates the basal heparin (Bio-Rad Laboratories, Richmond, Calif.) column level (30-ml bed volume). Protein was eluted from the heparin- of HIS4 expression. Furthermore, PHO2 was found to sepharose column by applying a Tris gradient. The pooled bind very strongly to the TRP4 promoter (G. Braus, H.-U. fractions contained approximately l0o PH04 and 2% Mosch, K. Vogel, A. Hinnen, and R. Hutter, EMBO J., in PHO2, respectively. Gel retardation assays. For gel retardation assays (10), the * Corresponding author. recessed 3' ends of the DNA fragments were filled in with 2050 VOL. 9, 1989 INTERACTIONS OF PHO2 AND PHO4 WITH THE PHOS GENE 2051

TATA -*Si//ATG +1 -541 -469 - 367 -245 -185 -100

ItI B B A H

A -O-400 32 (1) a 11 -, -36 (1w ) A12 -39-346 (0.1) A13 -350432 (0.6)

3.4 (0.4) 15- -2n (o.9)

16 -283-262 (l)

A 17 -255 174 (0.2)

A A8-22618 i2 -174, ( l) FIG. 1. The PHOS promoter. Negative numbers indicate elements found by deletion mapping (24). Listed beneath the sequence are some of the deletions used. Numbers in brackets show activities of the deletions in vivo. Restriction sites: A, AatII; B, BamHI; BE, BstEII; C, ClaI; H, HinPI; M, MnlI; S, Sail.

Klenow polymerase as described by Maniatis et al. (17), counts) was added to 38 ,ul (final volume) of reaction mixture using a-35S-labeled deoxynucleoside triphosphates. Approx- containing 50 mM NaCl, 10 mM Tris (pH 7.5), 2 mM imately 1 ng of labeled fragments was added to a 20-,ul (final dithiothreitol, 5 mM MgCl2, 0.1 mM EDTA, and 3 ,ug of volume) reaction mixture containing 100 mM NaCl, 10 mM double-stranded poly(dI-dC) competitor DNA. Then 1 ,ug of Tris (pH 7.5), 2 mM dithiothreitol, 0.1 mM EDTA, and 2 ,ug partially purified PHO4 protein was added. After the binding of double-stranded poly(dI-dC) competitor DNA. Finally, reaction was allowed to proceed for 15 min at 25°C, 2 RI (12.5 0.2 ,ug of PHO4 or 2 ,ug of PHO2 protein (partially purified) U) of exonuclease III (Boehringer Mannheim Biochemicals, was added, and the binding reaction was allowed to proceed Indianapolis, Ind.) was added for 3, 6, or 12 min at 25°C. The for 15 min at 25°C. The samples were loaded immediately reaction was stopped by addition of 4.8 of 500 mM Tris (under voltage) onto a 6% polyacrylamide gel (acrylamide- (pH 8.0)-S50 mM EDTA-5% sodium dodecyl sulfate and 2 bisacrylamide weight ratio of 30:1) made in 0.5x TBE. (50 of proteinase K (10 mg/ml). After proteinase K treatment, mM Tris [pH 8.3], 50 mM boric acid, and 1.25 mM EDTA). the samples were loaded onto a standard sequencing gel. Electrophoresis was performed at 350 V for 2 to 3 h at room temperature. The dried gels were exposed to sensitive X-ray RESULTS films (Eastman Kodak Co., Rochester, N.Y.). DNase I footprinting with PHO4. DNase I footprint assays PH04 and PH02 proteins bind specificafly to the PHOS were adapted from the method of Galas and Schmitz (9). A promoter. The promoter region of PH05 was screened for 32P-end-labeled DNA fragment (approximately 20,000 functionally important DNA sequences by the construction Cerenkov counts) was added to 20 ,ul (final volume) of of a series of deletion mutants (24) (Fig. 1). To determine reaction mixture containing 50 mM NaCl, 10 mM Tris (pH whether the regulatory proteins PHO2 or PHO4 bound 7.5), 2 mM dithiothreitol, 1 mM EDTA, 0.5 mM MgCl2, and directly to the PHOS promoter, we decided to analyze this 1.5 ,ug of poly(dI-dC) competitor DNA. After addition of 0.5 possible protein-DNA interaction in vitro. Both proteins ,ug of partially purified PHO4 protein, the binding reaction were expressed in E. coli (see Materials and Methods), was allowed to proceed for 15 min at 25°C. The samples were which eliminated the possibility of contamination by other transferred on ice, and 2 ,ul of DNase I (50 p.g/ml) freshly yeast proteins. After partial purification of the expressed diluted in 10 mM Tris (pH 7.5)-25 mM CaCl2 was added to proteins, gel retardation experiments were performed with the reaction mixtures. Digestion with DNase I was allowed PH05 promoter DNA. Radiolabeled DNA fragments were to proceed for 3 min on ice. The samples were loaded on a incubated with the regulatory proteins and subjected to standard sequencing gel. native polyacrylamide gel electrophoresis as described in DNase I footprinting with PHO2. DNase I footprint assays Materials and Methods. The presence of the PHO4 protein were done as described for PHO4 except that protein-DNA changed the mobility of the wild-type promoter fragment complexes were prepared with 5 ,ug of partially purified (Fig. 2). Distinct complexes migrating with a lower mobility PHO2 protein in the presence of 100 mM NaCl and the than that of the free DNA were observed. The interaction reaction volume was scaled up to 50 ,ul, containing approx- with PHOS promoter DNA was specific, since no binding imately 200,000 to 400,000 Cerenkov counts per sample. was observed with a control DNA fragment (Fig. 2, lane c) After DNase I treatment, 10 ,ul of 0.2 M EDTA (pH 7.5) was representing part of the PHOS coding region. added. The samples were loaded immediately onto a 6% The observed complexes appeared as two sets of triplet native polyacrylamide gel as described for gel retardation bands (Fig. 2). The individual bands within each triplet set assays. The wet gels were exposed to X-ray films, and the consistently- showed decreasing intensity with increasing two protein-specific complexes were isolated from the gel mobility, which suggested different stabilities of the various and loaded onto a standard sequencing gel. complexes. Figure 2 also shows protein-DNA complexes Exonuclease m footprinting. For protein-DNA binding, a with promoter fragments that harbored internal deletions (C7 32P-end-labeled fragment (approximately 20,000 Cerenkov to A20; Fig. 1). Most of these produced the same pattem as 2052 VOGEL ET AL. MOL. CELL. BIOL. wt C 7 8 9 10 11 12 13 14 15 16 171819 20 wt c 7 8 9 10 11 1213 14 15 16 17 18 19 20

_-. t- .

6" 6mw LdLAinL

FIG. 3. Localization of PH02-binding sites to the PH05 pro- FIG. 2. Localization of PHO4-binding sites to the PHOS promoter. Assay conditions and fragments were as for Fig. 2. The moter. Gel retardation assays were performed as described in fastest-moving bands represent free DNA, the middle band is Materials and Methods with I'S-labeled BamHI-BstEII promoter complex 1, and the slowest band is complex 2. fragments. The deletions (AlO to A18) are shown in Fig. 1. wt, Wild-type PHOS promoter fragment; c, a 370-base-pair fragment Two regions, from -347 to -373 (UASpl) and from -245 from the PHOS coding region used as a negative control. The to -262 (UASp2) (see Fig. 6A and C), were protected by the fastest-moving bands represent free DNA fragments. PH04 protein against DNase I digestion (Fig. 4). Deletions A11, A12, and A17 deleted DNA sequences from -362 to did the wild-type fragment, which indicated that a particular -392, -346 to -369, and -174 to -255, respectively (24), deletion did not interfere with PH04 binding. However, All, and overlapped perfectly with the two protected segments. A12, and A17 showed only one set of triplet bands, which DNase I footprint analysis using the BamHI-BstEII or suggested that one binding area for the PH04 protein was AatII-MnlI PHOS promoter DNA fragment (Fig. 1) showed missing. This finding coincides with the in vivo data, which no protected regions different from UASpl and UASp2 (data showed that All, A12, and A17 led to a reduced level of not shown), which indicated that the elements at -469 and PH05 mRNA (24). The observed triplet pattern could also -185 are not involved in PH04 binding even though they fit be explained as containing partially degraded PH04 protein. the consensus sequence (24) quite well. Our data argue However, the regular spacing of the single bands would strongly that PH04-mediated activation of the PH05 gene imply a specific kind of degradation, which is unlikely. occurs only via the two PH04-binding sites UASp1 and It is interesting that deletion A7, missing the element at UASp2. Furthermore, we could not detect by gel retardation -469, did not influence the gel retardation pattern. This assays and footprints any influence on PH04 binding of the finding is in complete agreement with our in vivo data presence of ATP, cAMP, or Pi in the binding reaction (data showing that deletion A7 produced no decrease in PHOS not shown). mRNA expression. Thus, the element at -469 must be Determination of the exact boundaries of the PH04-binding considered a fortuitous sequence with no biological signifi- sites by exonuclease m digestion. Exonuclease III footprint- cance. ing revealed a more precise footprint than did DNase I. In After the same gel retardation assay was done with brief, 5'-radiolabeled DNA fragments were incubated for partially purified PH02 protein, a significantly different different periods of time with exonuclease III in the presence pattern of complexes was obtained. The PH02 protein and absence of the PH04 protein. Stable protein-DNA yielded a double band on a native polyacrylamide gel (Fig. complexes do not allow exonuclease III to digest DNA 3). As was observed with the PH04 protein, the binding of within and downstream of the complex. This interference PH02 was specific for PHOS UAS DNA and did not occur will produce (radioactive) DNA fragments of particular with the control fragment. Fragments A15 and A16 did not lengths that can be resolved on sequencing gels. Individual bind PH02 protein, as judged by the absence of either exonuclease III digests from either end of a fragment define complex. This region obviously contains a sequence in- the left and right borders of a binding site involved in volved in PH02 binding. Unexpectedly, the deletions A15 protein-DNA interaction. and A16 showed only minor or no effects on PHOS mRNA Figure 5 shows the pattern obtained upon digestion of the expression in vivo. AatII-BstEII fragment unidirectionally by exonuclease III, PH04 protein protects two regions against DNase I. The starting at the BstEII site (Fig. 5A). In the absence of the PH04-binding sites defined by deletions All, A12, and A17 PH04 protein, the fragment became gradually degraded with were more precisely analyzed by DNase I protection exper- increasing incubation time. Some preferential stops that iments (9) as described in Materials and Methods. In brief, might be explained by colliding exonuclease III molecules the AatII-BstEII fragment of PHOS promoter DNA (Fig. 1) appeared after 6 min of digestion. However, in the presence was radioactively labeled at either 5' end and incubated with of the PH04 protein, a very strong exonuclease III stop was the PH04 protein. Protein-DNA complexes were treated observed at -245, the site of UASp2. In particular, it was briefly with DNase I. Proteins were subsequently digested interesting to note that no exonuclease III stop at UASpl, with proteinase K, and the DNA was electrophoresed on a which would result in a specific radioactive fragment with standard sequencing gel. higher electrophoretic mobility, was observed. VOL. 9, 1989 INTERACTIONS OF PHO2 AND PHO4 WITH THE PH05 GENE 2053

Lower specific exonuclease III stop at -368 that disappeared upon Upper longer digestion. After 12 min of digestion, however, a St rand Strand prominent stop occurred at -260, the UASp2 site. 0 0 The fact that no further digestion at -245 (UASp2) was -+ ,. obtained even after 12 min of incubation, whereas free DNA was almost completely digested, argues that the two sites behave in a distinctive manner with respect to PH04 binding. This observation could be confirmed by DNase I footprinting, using various amounts of partially purified PH04 protein. At low PH04 concentrations, UASp1 was barely -231 protected whereas UASp2 was fully protected (data not -373 shown). Figure 6A summarizes the protected sequences obtained 4( by DNase I and exonuclease III footprinting. The two -212 sequences showed little homology to each other except for -341 the pentamer CACGT, in approximately the middle of each i.a UASp. IL PH02 protein binds to a region between the PH04-binding sites. Having shown by gel retardation experiments (see above) that the PH02 protein binds specifically to a region between UASp1 and UASp2, resulting in two characteristic complexes, we became more interested in the nature ofthese complexes. We therefore combined gel retardation and DNase I footprinting to study the complexes in more detail. . ~- AatII-BstEII restriction fragments were labeled at either end and treated with DNase I in the presence or absence of the PH02 protein. Free DNA and the two complexes were resolved and recovered from a native polyacrylamide gel. S§# DNA was then subjected to denaturing polyacrylamide gel -r47 electrophoresis (see Materials and Methods). U§1 DNA in com- In comparison with free DNA, the present plex 1 (the faster-moving protein-DNA complex) showed a protein-DNA interaction between -277 and -2% (Fig. 7B), -212 recognized as a pattern of bands of different intensities. I. Unlike the footprint obtained with the PH04 protein, PH02 interaction also resulted in hypersensitive sites. n The same DNA region was involved in the protein-DNA -371L _ , interaction resulting in the formation of complex 2, the *_ _ _4e slower-moving complex. This finding indicates that PH02 mg ob Goo does not bind as a single molecule but rather forms higher- order complexes such as dimers or tetramers. This finding -241 was confirmed by the observation that A15, which removed the PH02-binding site, led to the disappearance of the two I complexes (Fig. 3). Also, A16 showed no PH02 binding (Fig. 3) and therefore must contain sequences important for protein-DNA interaction. The additional changes in intensity a *l Ii. of some bands, seen in complex 2 (Fig. 7B), could be _o ;L explained by higher-order structures of PH02; alternatively, FIG. 4. DNase I footprint analysis with the PHO4 protein. PH02 may bind to a second binding site, since we have some DNase I footprinting was performed as described in Materials and indications that PH02 can interact with adjacent sequences Methods. The upper strand was labeled at the AatII site, and DNA partially overlapping UASp2. was restricted with BstEII to generate the 270-base-pair fragment Obviously, the protein-DNA interaction for PH02 is much containing UASp1 and UASp2. The lower strand was labeled at the BstEII site and restricted with AatII. The DNA probes were weaker than that for PHO4. We have not been able to incubated in the presence of poly(dI-dC) competitor DNA with (+) stabilize the complexes by adding ATP, cAMP, Pi, or zinc or without (-) the PHO4 protein for 15 min and treated as described ions to the binding reaction. Furthermore, the presence of in Materials and Methods. A+G, Maxam-Gilbert sequencing reac- the PH04 protein in the binding reaction did not improve the tions (18) of the same DNA fragments used as size markers. PH02-DNA interaction, which suggests that PH04 and PH02 bind independently to the two UAS elements of the PH05 promoter. On the other hand, exonuclease III digestion starting at the ClaI site (Fig. SB) resulted in a specific stop at -349, the DISCUSSION UASp1 site (note that the AatII:ClaI fragment lacked UASp2). Furthermore, the digestion pattern of the comple- Our protein-DNA interaction studies have shown that mentary strand generated by labeling of the BstEII site and PH02 and PH04 bind specifically to the PHOS promoter digestion from the HinPI site (Fig. SC) showed two promi- DNA. Gel retardation experiments showed that two regions nent exonuclease III stops. A short digestion resulted in a on the PHOS promoter, defined by A11-A12 and A17 (Fig. 2054 VOGEL ET AL. MOL. CELL. BIOL.

A Upper Strand B Upper Strand C Lower Strand __ + _+ _EE +E+ a .5f m.2.9 CO IUD (7 C 9 Cs2(, a Ea 2 E E Nc * aE u U FE2 'I CDeN M (D 04 + CD C-4 + 01r *- - - 0At

.' .t.

(3.~..C, j j.~ ~j24 ....4 ..... ~. ... :. .:..

_0 '__. - 368 I ~ --245 0

__ _

w :^

4EWI,w S - - ,w Sp Wa.. a V. If * ' 0w 41S

4 w* _m 40 r. o

- *. _ -- 260 _ w .W

FIG. 5. Exonuclease III footprint analysis with the PH04 protein. Exonuclease III footprinting was performed as described in Materials and Methods. The upper strand was labeled at the AatII site and digested with exonuclease III from the BstEII site (A) or from the ClaI site (B). The AatII-BstEII fragment contains UASp1 and UASp2, whereas the AauII-ClaI fragment contains only UASp1. The lower strand was labeled at the BstEII site and exonuclease III digested from the HinPI site (C). The DNA probes were incubated in the presence poly(dI-dC) competitor DNA with (+) or without (-) the PH04 protein for 15 min and digested with 12.5 U of exonuclease III per assay for 3, 6, and 12 min. A+G, Maxam-Gilbert sequencing reactions (18) used as size markers.

6C), are involved in PHO4 binding and that another region, become independent of PHO2, since the PHO4 protein now defined by A15 and A16, is involved in PHO2 binding. directly interacts without need for the PHO2 protein. How- Footprint analysis mapped the PH04-binding sites at -347 ever, since the two deletion mutants A15 and A16 still to -373 (UASpl) and at -239 to -262 (UASp2) (Fig. 6A and depend on PHO2 in vivo (data not shown), this latter C), and exonuclease III stops defined binding regions at explanation is impossible. -349 to -368 and at -245 to -260. These PHO4-binding It was recently shown that PH02 is involved not only in sites are in good agreement with the phosphate control phosphatase gene regulation but also in regulation of the sequences defined previously (24). basal level of HIS4 (3). The authors gave an approximate The PH02 protein was found to occupy one region, localization of the PH02 (BAS2)-binding site on the PH05 between -277 and -296 (Fig. 6B and C). The footprint data promoter by using the gel retardation assay. Moreover, it for PH02 showed that complex 1 as well as complex 2 (Fig. was shown that PH02 is involved in the basal expression of 3) protected approximately the same area (Fig. 7). The CYCI (26). Furthermore, we have shown that PH02 binds different intensities of some additional bands in complex 2 specifically to the TRP4 promoter (Braus et al., in press). In (lower strand) could have resulted from the higher-complex the latter case, PH02 is not essential for basal level expres- formation of the PHO2 protein. More interesting, we have sion of TRP4 but clearly modulates the general control some indications that PH02 may bind in vitro to an adjacent response of TRP4. Interestingly, the in vitro binding of binding site that slightly overlaps UASp2 (data not shown). PHO2 to the TRP4 promoter is much stronger than the Deletions in the PHO2-binding region (M15 and A16) binding to the PHOS promoter, resulting in a very clear showed no or only a minor decrease of PH05 mRNA footprint pattern. expression in vivo (24). The same deletions, however, Unlike the PH02 protein, PHO4 seems to be a specific showed a drastic effect in vitro, since PHO2 binding was regulator of the repressible phosphatase genes. Comparisons completely abolished. An explanation for these conflicting between UASp1 and UASp2 revealed a short sequence, results is that because ofthe shortened distance between A15 CACGT, common to both UASs. Furthermore, this se- and A16 in the two PH04-binding sites, gene activation may quence is also found in the PHOIJ (24) and PH08 (12) VOL. 9, 1989 INTERACTIONS OF PHO2 AND PHO4 WITH THE PH05 GENE 2055

A PH04 binding sites A Upper Strand B Lower Strand a 0 f1F F2 a FI F2 F + -370I_ UApUASpl -347 F0 aata t a t TAATTAGCACGTTTTCGCATA GAacgc t tatATAAT TTAATCGTGCAAAAGCGTAT c t tgcg "Iiiti"i^455 - 373 - 349

-215 - 262 -239 I. u UASp 2 - _. ac c tTCACTCACACGTGGGACTAGCAcaga -29 tggaA CTGAGTGTGCACCCTGATCg t g tct - - -262 -241 a_b-

I7-uI-Xe -m~as3W.m B PH02 binding site -21-w- - -296 -279 0<-. -.-_- -277L7 7-m__t aat t GAATAGGCAATCTCTAAAt gaatc _211_ot ,- n t taaCTTATCCGTTAGAGATTTAC t tag I -296 -277 |_S_3S_8_ C UASpl UAS,2 FIG. 7. DNase I footprint analysis with PHO2. DNase I foot- '_ % TATA ATO printing was performed as described in Materials and Methods. The 5~~- 6~~~ -E3 - I141 I9 I -367 -245 l upper and lower strands of the AatII-BstEII restriction fragment were labeled as described in the legend to Fig. 4. The DNA probes were incubated with the PH02 protein, digested with DNase I, and FIG. 6. Summary of PHO4- and PH02-binding sites defined by electrophoresed on a native polyacrylamide gel. The free DNA DNase I and exonuclease III footprinting. (A) PHO4-binding sites. band, complex 1, and complex 2 were isolated, and the DNA was Sequences protected against DNase I digestion are indicated by bold analyzed on a standard sequencing gel. A+G, Maxam-Gilbert capital letters; exonuclease III footprints are boxed. (B) PH02- sequencing reactions (18) used as size markers. Protected regions binding site. The DNase I-protected region is shown in bold capital are bracketed on the left of each panel. letters. (C) Arrangement of UASp1, UASp2, and the PH02-binding site on the PH05 promoter. Restriction sites are as described in the legend to Fig. 1. For some deletions used by Bergman et al. (6), the decreased activity appears to have been directly related to the lengths promoter regions. All of these genes depend on PHO4 for of the deletions. Furthermore, their studies were all done on activation. It is most likely that this sequence corresponds to multicopy plasmids, whereas the deletion studies by Ru- the core sequence of the PHO4-binding element. dolph and Hinnen (24) were done on integrated gene copies. Studies of the regulation of CYCI have shown that binding The PHO4 data presented here are in good agreement with of HAP1 in vitro, like the activity of UAS1 in vivo, is the data presented by Rudolph and Hinnen (24) (summarized stimulated by heme (21, 22). GAL4, on the other hand in Fig. 8). Both All and A12, which caused a 10-fold requires zinc for binding to DNA in vitro and in vivo (11). decrease of PHOS mRNA, affect part of the PHO4-binding We therefore asked whether PHO4 or PHO2 binding is sequence determined by footprinting. Also, A17, which influenced by certain low-molecular-weight factors and in- showed a fivefold decrease in PHOS mRNA (24), deletes vestigated the binding behavior in the presence of Pi, ATP, important sequences for PHO4 binding. cAMP, and zinc ions. None of these components produced In comparisons of our data with the chromatin structure of any changes in binding properties (data not shown). the PHOS promoter (2), it is interesting that UASp1 coin- We have compared different in vivo data, obtained by cides with a hypersensitive region that is observed indepen- deletion mapping (7, 19, 24), with our in vitro data. The dent of gene activation. This finding suggests that UASp1 is UASJ (-382 to -391) and UASI (-332 to -341) proposed not protected by a nucleosome and may always be available by Bergman et al. (6) do not coincide with our UASs for protein-DNA interactions. Also, it is conceivable that the obtained by DNase I and exonuclease III footprinting (Fig. PHO4 protein is not the only regulator protein interacting 6A). Also, UAS II (-301 to -340) proposed by Nakao et al. with UASp1; it is possible that PHO80, the negative factor (19) does not coincide with our UASs. However, the UAS I for the repressible PH05 gene, binds to the same or an (-354 to -390) partially overlaps our UASp1 (-347 to overlapping sequence, as has recently been suggested by -373). Neither group found the element at -239 to -262 Madden et al. (16). In comparison, experiments with the (UASp2) (Fig. 6A and C), which we show to be important for CYCI promoter have shown that the HAP] activator not both PH05 activation in vivo and binding in vitro (24; this only requires heme for optimal binding to UAS1 but also study). These divergent results could be attributable to the competes with the factor RC2 for the same binding site (21). fact that Bergman et al. (6) and Nakao et al. (19) used quite The interaction of the PHO2 protein with the PHOS large deletions for the localization of regulatory elements. promoter cannot be explained readily, since the PHO2 2056 VOGEL ET AL. MOL. CELL. BIOL.

UASpl - 392 -368 -361 -346 A11* GTCTGCACAAAGAAATATATATTAAATTAGCA A 12* AAATTAGCACGTTTTCGCATAGAA - 367 element * ATATATTAAATTAGCACGT DNAase I TATTAAATTAGCACGTTTTCGCATAGA ExofM AATTAGCACGTTTTCGCATA UASp2 - 262 -254 - 245 -225 -174

A 17* CACACGTGGGACTAGCACAGACTAAATTTAT .. AATTG - 245 element * ACACGTGGGACTAGCACAG DNA ase I TGGCACTCACACGTGGGACTAGCA ExoM GCACTCACACGTGGGA FIG. 8. Comparison between in vivo and in vitro data. *, Data published by Rudolph and Hinnen (24). All, A12, and A17 show sequences that are missing in the respective deletion derivatives of the PH05 gene. The -367 and -245 elements were deduced from deletion mapping. DNase I and Exo III, Protected regions in the respective footprint analyses. Sequences coinciding among the different experiments are underlined.

protein binds only weakly to the PH05 promoter in vitro. In chromatin structure associated with derepression of the acid addition, deletions of this binding site show no or little phosphatase gene of Saccharomyces cerevisiae. J. Biol. Chem. influence in vivo on PH05 expression (24). These observa- 258:7223-7227. tions leave open the possibility that the PH02 protein is 6. Bergman, L. W., D. C. McClinton, S. L. Madden, and L. H. involved in the synthesis of some other regulatory proteins. Preis. 1986. Molecular analysis of the DNA sequences involved in the transcriptional regulation of the phosphate-repressible Alternatively, it could exert its function by contact with acid phosphatase gene (PHO5) of Saccharomyces cerevisiae. other proteins of the PH05 activation machinery. Proc. Natl. Acad. Sci. USA 83:6070-6074. Recent findings revealed that PH02 encodes a homeobox 7. Bergman, L. W., M. C. Stranathan, and L. H. Preis. 1986. (8). Very likely, PH02 is a more generally acting regulatory Structure of the transcriptionally repressed phosphate-repress- protein that controls the activation of unrelated genes. In the ible acid phosphatase gene (PH05) of Saccharomyces cerevi- case of PHO2, it has been demonstrated that, in addition to siae. Mol. Cell. Biol. 6:38-46. some genes involved in phosphate metabolism (PHO5, 8. Burglin, T. R. 1988. The yeast regulatory gene PHO2 encodes a PHOIJ, and PHO84), HIS4, CYCI, and TRP4 (3, 24; Braus homeo box. Cell 53:339-340. et al., in press) depend on the action of PHO2. Taken 9. Galas, D. J., and A. Schmitz. 1978. DNase footprinting: a simple together, the data available point to an interesting dual method for the detection of protein-DNA binding specificity. function of the PH02 protein, i.e., interaction with promoter Nucleic Acids Res. 5:3157-3170. 10. Garner, M. M., and A. Revzin. 1981. A gel electrophoresis DNA as well as with other proteins ofthe activation process. method for quantifying the binding of proteins to specific DNA The apparent difference in the binding affinities of PH02 for regions: application to components of the Escherichia coli the respective promoters may indicate a distinct balance of lactose operon regulatory system. Nucleic Acids Res. 9:3047- protein-DNA and protein-protein contacts for different sys- 3060. tems. 11. Johnston, M. 1987. Genetic evidence that zinc is an essential co-factor in the DNA binding domain of GAL4 protein. Nature ACKNOWLEDGMENTS (London) 328:353-355. 12. Kaneko, Y., N. Hayashi, A. Toh-e, I. Banno, and Y. Oshima. We thank K. A. Bostian for providing the PH04 clone, C. 1987. Structural characteristics of the PH08 gene encoding Sengstag for helpful discussions and critical reading of the manu- repressible alkaline phosphatase in Saccharomyces cerevisiae. script, and A. Schmitz, B. Meyhack, and H. Hirsch for fruitful Gene 58:137-148. discussions. 13. Koren, R., J. A. LeVitre, and K. A. Bostian. 1986. Isolation of the positive-acting regulatory gene PHO4 from Saccharomyces LITERATURE CITED cerevisiae. Gene 41:271-280. 1. Almer, A., and W. Horz. 1986. Nuclease hypersensitive regions 14. Kunkel, T. A. 1985. Rapid and efficient site-specific mutagenesis with adjacent positioned nucleosomes mark the gene boundaries without phenotypic selection. Proc. Natl. Acad. Sci. USA of the PHO5/PHO3 locus in yeast. EMBO J. 5:2681-2687. 82:488-492. 2. Almer, A., H. Rudolph, A. Hinnen, and W. Horz. 1986. Removal 15. Legrain, M., M. De Wilde, and F. Hilger. 1986. Isolation, of positioned nucleosomes from the yeast PHO5 promoter upon physical characterization and expression analysis of the Sac- PHO5 induction releases additional upstream activating DNA charomyces cerevisiae positive regulatory gene PHO4. Nucleic elements. EMBO J. 5:2689-2696. Acids Res. 14:3059-3071. 3. Arndt, K. T., C. Styles, and G. R. Fink. 1987. Multiple global 16. Madden, S. L., C. L. Creasy, V. Srinivas, W. Fawcett, and L. W. regulators control HIS4 transcription in yeast. Science 237: Bergman. 1988. Structure and expression of the PHO80 gene of 874-880. Saccharomyces cerevisiae. Nucleic Acids Res. 16:2625-2637. 4. Baiwa, W., B. Meyhack, H. Rudolph, A. M. Schweingruber, and 17. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular A. Hinnen. 1984. Structural analysis of the two tandemly cloning: a laboratory manual. Cold Spring Harbor Laboratory, repeated acid phosphatase genes in yeast. Nucleic Acids Res. Cold Spring Harbor, N.Y. 12:7721-7739. 18. Maxam, A. M., and W. Gilbert. 1980. Sequencing end-labelled 5. Bergman, L. W., and R. A. Kramer. 1983. Modulation of DNA with base-specific chemical cleavages. Methods Enzymol. VOL. 9, 1989 INTERACTIONS OF PHO2 AND PHO4 WITH THE PH05 GENE 2057

65:499-560. activator binds to two upstream activation sites of different 19. Nakao, J., A. Miyanohara, A. Toh-e, and K. Matsubara. 1986. sequence. Cell 49:19-27. Saccharomyces cerevisiae PH05 promoter region: location and 23. Remaut, E., P. Stanssens, and W. Fiers. 1981. Plasmid vectors function of the upstream activation site. Mol. Cell. Biol. 6: for high-efficiency expression controlled by the PL promoter of 2613-2623. coliphage lambda. Gene 15:81-93. 20. Parent, S. A., A. G. Tait-Kamradt, J. LeVitre, 0. Lifanova, and 24. Rudolph, H., and A. Hinnen. 1987. The yeast PHO5 promoter: K. A. Bostian. 1987. Regulation of the phosphatase multigene phosphate-control elements and sequences mediating mRNA family of Saccharomyces cerevisiae, p. 63-70. In A. Torriani- start-site selection. Proc. Natl. Acad. Sci. USA 84:1340-1344. Gorini, F. G. Rothman, S. Silver, A. Wright, and E. Yagil (ed.), 25. Sengstag, C., and A. Hinnen. 1987. The sequence of the Sac- Phosphate metabolism and cellular regulation in microorgan- charomyces cerevisiae gene PHO2 codes for a regulatory pro- isms. American Society for Microbiology, Washington, D.C. tein with unusual aminoacid composition. Nucleic Acids Res. 21. Pfeifer, K., B. Arcangioli, and L. Guarente. 1987. Yeast HAP1 15:233-246. activator competes with the factor RC2 for binding to the 26. Sengstag, C., and A. Hinnen. 1988. A 28 bp segment of the S. upstream activation site UAS1 of the CYCl gene. Cell 49:9-18. cerevisiae PHO5 UAS confers phosphate control to the heter- 22. Pfeifer, K., T. Prezant, and L. Guarente. 1987. Yeast HAP1 ologous gene CYC1-lacZ. Gene 67:223-228.